Problems with Vex Pro mecanum wheels

[Calvin Hartley]

We used them this year and were very satisfied. Of course it will be slower sideways, but I don't recall exactly how slow it is for us. (We also used 1 cim to each wheel, but I don't remember the gear ratio)

[omsahmad]

Originally Posted by Calvin Hartley

We used them this year and were very satisfied. Of course it will be slower sideways, but I don't recall exactly how slow it is for us. (We also used 1 cim to each wheel, but I don't remember the gear ratio)

The robot with our homemade wheels is also certainly noticeably slower sideways vs. forwards/backwards, but as slow as the vex wheels. I can't remember how the andymark wheels performed exactly, but I do recall they were better than these vex wheels.

http://www.youtube.com/watch?v=W0XN6iSvCXo

^This is the kind of speed I am hoping for...

[JesseK]

While moving forward/back, your torque is the same as a normal wheel.

While strafing, your sideways torque is 1/SQRT(2) (~70%) of your normal torque since the wheels are driving at each other BEFORE the extra losses of friction in the rollers.

While on a diagonal, your acceleration is somewhere between 35-50% of your normal forwards acceleration since you're only spinning half of the motors and also half of your your robot is (essentially) on caster wheels.

Translating torque into acceleration has a lot to do with roller friction and robot weight. Additionally, unless there are specific dead zones for sideways & forward, the code is robbing the robot of torque needed to strafe at reasonable speeds (since no joystick is ever perfectly sideways for long).

[Ether]

Originally Posted by JesseK

While strafing, your sideways torque is 1/SQRT(2) (~70%) of your normal torque since the wheels are driving at each other BEFORE the extra losses of friction in the rollers.

What torque are you referring to in this context?

[JesseK]

Originally Posted by Ether

What torque are you referring to in this context?

Torque available for acceleration.

[Ether]

It's the reaction force of the floor on the wheel which causes acceleration of the vehicle.

In the ideal case of no roller friction or axial free play, and a non-compliant floor, the acceleration force in the strafing direction is the same as the accelerating force in the forward direction.

The affordable mecanum wheels used for FRC do not have thrust ball bearings on the rollers. The rollers have axial free play. The rollers have spinning friction (especially when loaded). The carpet is compliant. That's what makes the vehicle go slower in the strafing direction.

[JesseK]

Originally Posted by Ether

It's the reaction force of the floor on the wheel which causes acceleration of the vehicle.

In the ideal case of no roller friction or axial free play, and a non-compliant floor, the acceleration force in the strafing direction is the same as the accelerating force in the forward direction.

The affordable mecanum wheels used for FRC do not have thrust ball bearings on the rollers. The rollers have axial free play. The rollers have spinning friction (especially when loaded). The carpet is compliant. That's what makes the vehicle go slower in the strafing direction.

What do you mean by 'compliant'?

[Ether]

Originally Posted by JesseK

What do you mean by 'compliant'?

Definition 3.

[omsahmad]

Originally Posted by Ether

It's the reaction force of the floor on the wheel which causes acceleration of the vehicle.

In the ideal case of no roller friction or axial free play, and a non-compliant floor, the acceleration force in the strafing direction is the same as the accelerating force in the forward direction.

The affordable mecanum wheels used for FRC do not have thrust ball bearings on the rollers. The rollers have axial free play. The rollers have spinning friction (especially when loaded). The carpet is compliant. That's what makes the vehicle go slower in the strafing direction.

Ok that makes sense. We did dedue it had something to do with the spinning wheels.

Any recommendations to make it better?

[Ether]

Originally Posted by omsahmad

Any recommendations to make it better

Glue the rollers to their axles and forget about strafing.

[mr.roboto2826]

Originally Posted by Ether

Glue the rollers to their axles and forget about strafing.

Or this? http://www.chiefdelphi.com/media/photos/36998

[Mark Sheridan]

Make sure to check every roller, last year we had a strafing issue and it was because one roller was stuck. Our issue looked more like an intermittent speed issue is strafing.

I would also check if each wheel can rotate forward and backwards at full speed with the robot off the ground. Hopefully you have the encoders hooked up to each gear box to see the counts per second. At least this can sort out programing versus mechanical issues.

Just to be clear how loose our rollers are on our 2012 robot, the rollers will spin a couple time after flicking them.

[omsahmad]

Originally Posted by Mark Sheridan

Make sure to check every roller, last year we had a strafing issue and it was because one roller was stuck. Our issue looked more like an intermittent speed issue is strafing.

I would also check if each wheel can rotate forward and backwards at full speed with the robot off the ground. Hopefully you have the encoders hooked up to each gear box to see the counts per second. At least this can sort out programing versus mechanical issues.

Just to be clear how loose our rollers are on our 2012 robot, the rollers will spin a couple time after flicking them.

Ok thanks, I'll test all of this out Wednesday. I would prefer to keep mecanum for this year.

[Ether]

Originally Posted by Mark Sheridan

Just to be clear how loose our rollers are on our 2012 robot, the rollers will spin a couple time after flicking them.

Problem is, the rollers can appear to be free-spinning when not under load, and then bind up when they are loaded.

There is a lot of axial force on the roller under real-world operating conditions. It is prohibitively expensive to put high-quality thrust ball bearings on FRC mecanum wheel rollers. I've never seen the VEX mecanum wheels. What provision is in the design to react the high axial forces without creating excessive friction? IOW, what two materials are scrubbing against each other when the roller is axially loaded?

[Mark Sheridan]

Originally Posted by Ether

Problem is, the rollers can appear to be free-spinning when not under load, and then bind up when they are loaded.

There is a lot of axial force on the roller under real-world operating conditions. It is prohibitively expensive to put high-quality thrust ball bearings on FRC mecanum wheel rollers. I've never seen the VEX mecanum wheels. What provision is in the design to react the high axial forces without creating excessive friction? IOW, what two materials are scrubbing against each other when the roller is axially loaded?

Yeah I was going to caution that a wheel unloaded will be different from a loaded wheel. My team never came up with a good test method for this.

I am sure you know this Ether, but so others can follow, I will be a bit more detailed. On the andy mark wheels, there is a bushing in the wheel. This rotates on a bolt. Between the wheel and hub plates are washers. If one is too zealous, one can tighten the bolts too much compressing the washers and wheels until the wont move. We had to play with this a bit because once the wheels are on the ground, the deflection of parts can create binding too. I think if its too loose, the hub plates could touch the roller. I forget how much it took to cause the failure mode, but gut tells me that situation required the bolts of be barely tighten.

I am guessing the Vex wheels are similar with a shoulder bolt or screw that supports the rollers. Thus i may be possible to over tighten the screw until the rollers bind. The roller would be compressed by the head of the screw and the hub ( plus a few washers in there).